MesoNet: a Compact Facial Video Forgery Detection Network

TLDR

Traditional image forensics techniques are ill‑suited to videos because compression degrades the data. The study aims to automatically and efficiently detect face tampering in videos, focusing on Deepfake and Face2Face, by employing two lightweight deep‑learning networks that target mesoscopic image properties. The authors develop and evaluate these fast, low‑layer networks on an existing dataset and a newly constructed online‑video dataset. The tests achieve over 98 % accuracy for Deepfake and 95 % for Face2Face detection.

Abstract

This paper presents a method to automatically and efficiently detect face tampering in videos, and particularly focuses on two recent techniques used to generate hyper-realistic forged videos: Deepfake and Face2Face. Traditional image forensics techniques are usually not well suited to videos due to the compression that strongly degrades the data. Thus, this paper follows a deep learning approach and presents two networks, both with a low number of layers to focus on the mesoscopic properties of images. We evaluate those fast networks on both an existing dataset and a dataset we have constituted from online videos. The tests demonstrate a very successful detection rate with more than 98% for Deepfake and 95% for Face2Face.